dialects: (arith) Add support for bitcast operation

[jakedves]

Implementation of MLIR's arith.bitcast op. Test cases are passing but type checking on t1 and t2 variables isn't

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 91.27%. Comparing base (2338bb0) to head (d60ce88).
Report is 58 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #3805      +/-   ##
==========================================
+ Coverage   91.23%   91.27%   +0.03%     
==========================================
  Files         461      462       +1     
  Lines       57487    57719     +232     
  Branches     5548     5569      +21     
==========================================
+ Hits        52449    52683     +234     
+ Misses       3615     3614       -1     
+ Partials     1423     1422       -1     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[compor]

Very good, thanks for this. I've left a few comments.

[jakedves]

I've made all the changes suggested besides moving that function to xdsl.utils.type because it handles an edge case (IndexType) that I don't think would be useful/apply to other ops. Renamed it to make it clearer that its not generic if that's alright?

[compor]

A few nits!

[superlopuh]

@jakedves, as a convention, we tend to let the original commenter resolve the messages, it makes it easier to see how the change was addressed, and whether the conversation is really resolved.

[compor]

LGTM modulo a green CI run (the mlir-opt flag has a typo)

[superlopuh]

Thank you for this!